StudyBrixFree worksheets, study guides and courses

Artificial intelligence · University

Parameter count in a dense layer

Assume one trainable bias per output and no parameter sharing.

Make a worksheet

Method and conditions

Assume one trainable bias per output and no parameter sharing.

parameters = inputs×outputs + outputs

Worked examples

A fully connected layer maps 22 inputs to 22 outputs with one bias per output. How many trainable parameters does it have? Round the final result to 4 decimal places if needed.

There are 484 weights plus 22 biases.

506

A fully connected layer maps 56 inputs to 15 outputs with one bias per output. How many trainable parameters does it have? Round the final result to 4 decimal places if needed.

There are 840 weights plus 15 biases.

855

Try it

  1. A fully connected layer maps 21 inputs to 24 outputs with one bias per output. How many trainable parameters does it have? Round the final result to 4 decimal places if needed.

  2. A fully connected layer maps 50 inputs to 23 outputs with one bias per output. How many trainable parameters does it have? Round the final result to 4 decimal places if needed.

Check your work

Answers and marking notes
  1. 528

    There are 504 weights plus 24 biases.

  2. 1173

    There are 1150 weights plus 23 biases.

Further reading and review references

These are reference links, not copied exercises, institutional endorsement or proof of independent review.

AI-assisted preview; independent educational review required. Level labels describe prerequisites, not accreditation or official curriculum alignment.